How Google Search Is Trending in 2026: Reading the Signals

Most years, “how Google is trending” means the rankings moved. A core update lands, some sites go up, some go down, and we argue about why. 2026 has been different, because the product itself changed shape rather than just shuffling positions. The thing we have all been optimizing for is not really the same thing it was eighteen months ago.

So this is not a list of ranking factors. It is me reading the signals Google has been sending and saying what each one means for the work. If you want the wider picture this fits into, it sits under my pillar on SEO Trends 2026. Here I am just zooming in on Google itself.

The signal: AI Mode went from experiment to default

Start with the headline fact, because it is the one that reframes everything else.

AI Mode launched in the US in May 2025. By late 2025 it had rolled out to more than 180 countries. Within a year it passed a billion users, with queries more than doubling every quarter, per Google’s own numbers. Then at Google I/O 2026 on May 19, Google made the AI-first experience the default way Search works. Same event, they upgraded it to a new Gemini 3.5 Flash model, shipped a reimagined search box, and previewed background agents and agentic checkout.¹

A growth curve like that is not what a company shows you for a feature it is still deciding on. Google moved AI Mode from opt-in to default because the usage gave it cover to commit, and the commitment is the part that matters for the work.

For an SEO, the meaning is fairly simple. For a growing share of searches, the first thing a person meets on Google is a back-and-forth rather than a list of links, and that shows up as a change in behavior well before it shows up as a change in rankings.

The signal: the ten blue links are not gone

It is worth keeping the panic honest, though. Google’s VP of Search, Liz Reid, has said plainly that the ten blue links are not gone, and classic results still exist, still rank, and still get clicked.

That matters because the loud version of this story is that Google killed organic search, which is not what happened. What changed is which experience the user meets first. Classic results are still there, but for a growing share of sessions they sit one tap away inside a conversation rather than being the thing you land on.

In practice that does not mean abandoning classic SEO. It means classic SEO turns into the supply line for the AI experience instead of the experience itself, because the pages that rank are the pages the conversation pulls from. I traced how that relationship formed in how SEO changed from 2023 to 2026.

The signal: AI Overviews are on roughly half of searches

Separate from AI Mode, AI Overviews now appear on roughly half of all searches². This is the version most people have already seen, even if they never opened AI Mode on purpose.

The interpretation here is about attention. When an Overview sits at the top, the user reads an answer before they consider a link. Sometimes they never reach the links at all. The query was satisfied above the fold, by a synthesis that may or may not name your site.

So on these queries the job is no longer to rank in the top three and wait for the clicks to come. It is to be one of the sources the Overview gets built from, which rewards different things than chasing position alone. I broke the formatting and the trade-offs down in AI Overviews vs classic SEO. If you do not get cited, the ranking earns you less than it used to.

The signal: core updates keep getting more volatile

The last signal is the familiar one that keeps getting louder: core updates keep getting more volatile. The March 2026 update reshuffled about 80% of top-three results across tracked queries³, which is a lot of movement for positions that used to sit relatively still.

I do not read that as Google flailing so much as Google retuning what it trusts, faster and harder, because the answers it generates are only as good as the pages it pulls from. Once the model is doing the synthesis, the cost of trusting a weak page goes up, so the filter tightens and the swings get bigger.

The takeaway is the same one it has been, just with higher stakes attached. Pages with depth, first-hand experience, and trust signals tend to hold through these updates, while thin, templated pages get cleared out faster than they used to be.

What all four signals add up to

Put them together and the pattern is clear enough: search is turning into a conversation, where people ask longer, multi-step questions and then follow up inside the same session.

That carries a measurement consequence most teams have not absorbed yet. A single AI Mode session can cover six topics without ever loading a visible results page. The person got their answers and the demand was real, but your analytics see fewer sessions, because there were fewer page loads to count. Read that drop as falling demand and you will make the wrong call.

The job shifts with it. It used to be enough to rank for a keyword, and now the goal is to be the source the conversation pulls from. Ranking still feeds that, which is why classic SEO still matters, but it has become the input to the process rather than the finish line.

What I would actually do about it

Four moves, and none of them are exotic.

1. Optimize for question-and-follow-up journeys, not single keywords. Map the real sequence a person asks, the first question and then the obvious next two, and answer all of them on the page in that order. The conversation rewards pages that anticipate the follow-up.
2. Structure content for extraction. Lead with a tight, quotable answer in the first 60 words, use question-shaped headings, and add comparison tables and clear definitions. Give the model an easy sentence to lift and it will lift yours.
3. Expect more zero-click research, and plan for it. A lot of value now happens before any click, so build brand presence and consistency across the web. When the answer is assembled without a click, your name should still be in it.
4. Fix your measurement. Stop treating session count as a clean proxy for demand, and track whether you appear in AI answers and whether they represent you correctly. Falling sessions with steady visibility is a measurement artifact, not a business problem.

That is the year so far for Google itself. The interface keeps shifting, but what wins underneath it has barely changed. A genuinely useful page that a model can parse and actually trust still comes out ahead, the same as it did before any of this started. The shape of the product changed, but the work underneath it is still recognizable.

Frequently asked questions

Is Google AI Mode the default now?

Yes. At Google I/O 2026 on May 19, Google made the AI-first experience the default way Search works, running on a new Gemini 3.5 Flash model. AI Mode had already passed a billion users within a year of its May 2025 US launch, with queries more than doubling every quarter, which is why Google promoted it from opt-in to default.

Did Google get rid of the ten blue links?

No. Google's VP of Search, Liz Reid, said plainly that the ten blue links are not gone, and classic results still exist. What changed is the default experience. For a growing share of sessions the conversation comes first, and the familiar list of links sits one tap away rather than being the page you land on.

How does AI Mode change SEO?

It moves the goal from ranking for a keyword to being the source the conversation pulls from. People ask longer, multi-step questions and follow up in one session. Classic ranking still feeds the AI answer, so it matters, but it is now the input rather than the finish line. Format pages so a model can extract and cite them.

Do AI Overviews appear on most searches?

AI Overviews now appear on roughly half of all searches. On those queries, the user often reads a synthesized answer before considering any link, and sometimes never reaches the links at all. The practical job is to be one of the sources the Overview is built from, not just to hold a top-three ranking position.

Why are my Google sessions falling even though demand is steady?

Because a single AI Mode session can cover several topics without ever loading a visible results page, so there are fewer page loads to count. Demand can hold steady while session counts drop. Read that as a measurement artifact, not falling demand, and track whether you appear in AI answers rather than relying on session volume alone.

Sources

1. Google AI Mode passed one billion monthly users and became the default Search experience on the new Gemini 3.5 Flash model. Google, Search at I/O 2026. ↩

2. AI Overviews appeared on roughly 48% of tracked queries by February 2026. BrightEdge, 2026. ↩

3. In the March 2026 core update, 79.5% of top-3 URLs changed positions, more volatile than December 2025’s 66.8%. Search Engine Land, 2026. ↩